Repolode
Submit
HumanSignal

HumanSignal/label-studio

Open source data labeling tool for annotating audio, text, images, video and time series, with exports to ML model formats.

Open-source alternative to:LabelboxScale AI

View on GitHubWebsite
Stars
28,647
Forks
3,787
Open issues
957
Last commit
5 hours ago

What is it?

Label Studio is an open source tool for labeling raw data such as audio, text, images, videos and time series. It provides a configurable web interface where teams can define label formats and export annotations to various model formats. It is used to prepare new datasets or improve existing training data for machine learning models.

The tool supports multiple users with accounts, multiple projects per instance, and imports from local files or cloud storage such as Amazon S3 and Google Cloud Storage. A REST API allows labeling to be built into a data pipeline. Model backends can be connected to pre-label data, compare predictions, or run active learning on the most uncertain examples.

Label Studio can be run with pip, Docker, Docker Compose with PostgreSQL and Nginx, or deployed to cloud platforms. The project is distributed under the Apache 2.0 license and includes a frontend library and a data manager library as separate components.

Key features

  • 01Labels audio, text, images, video, HTML and time-series data.
  • 02Configurable label interface defined by a templating configuration language.
  • 03Multi-user accounts and multiple projects in one instance.
  • 04Imports from local files, AWS S3 and Google Cloud Storage.
  • 05REST API for embedding labeling into data pipelines.
  • 06Connects ML backends for pre-labeling and active learning.

Use cases

  • Creating labeled training datasets for computer vision models such as object detection or segmentation.
  • Annotating text for NLP tasks like classification or named entity recognition.
  • Reviewing and correcting model predictions to improve existing training data.
  • Running a shared labeling workspace for a team with multiple annotators.

Quick start

  1. 1

    Install with pip (requires Python 3.10+)

    pip install label-studio
  2. 2

    Start the server, then open http://localhost:8080

    label-studio

Things to know

  • Local pip install requires Python 3.10 or newer.
  • Windows setup may need manually downloaded lxml wheels from Gohlke builds.

In collections

Computer vision libraries & toolsNatural language processing tools & libraries

Topics

#annotation#data-labeling#computer-vision#machine-learning#image-annotation#nlp#mlops#dataset#annotation-tool#annotations

FAQ

What is label-studio?
Open source data labeling tool for annotating audio, text, images, video and time series, with exports to ML model formats. Label Studio is an open source tool for labeling raw data such as audio, text, images, videos and time series. It provides a configurable web interface where teams can define label formats and export annotations to various model formats. It is used to prepare new datasets or improve existing training data for machine learning models.
Is label-studio free to use? What license does it have?
label-studio is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is label-studio still maintained?
Actively maintained. The last commit to label-studio was on Nov 7, 2026.
What can I use label-studio for?
Creating labeled training datasets for computer vision models such as object detection or segmentation. Annotating text for NLP tasks like classification or named entity recognition. Reviewing and correcting model predictions to improve existing training data. Running a shared labeling workspace for a team with multiple annotators.
How do I install label-studio?
Install with pip (requires Python 3.10+): pip install label-studio Start the server, then open http://localhost:8080: label-studio
Where can I get label-studio?
The source code and installation instructions are on GitHub: https://github.com/HumanSignal/label-studio

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects